1) I've used several brands of FSRs, including FlexiForce and you're very unlikely to get results usable in trade; their responses to pressure vary with temperature (because, I suspect, the stiffness of the polymers in the sensor itself vary w/temp.)
2) The application guides I've read strongly suggest NOT using extremely stiff materials or sharp radii in direct contact with the sensor that can cause stress concentrations.
3) The stiffness and shape of both what's on top of the sensor and what's underneath both matter.
4) all of the materials you're considering are orders of magnitude stiffer than the sensor, so they'll likely give you equally wrong and expensive results.
5) I suggest you first try either thin disks or very-slightly rounded domes made with something like Nylon or HDPE (High-density polyethylene, AKA cutting-board plastic.) I suspect you can back those with disks of a more rigid material; steel would be fine.
6) The disk has to be same-size or slightly smaller than the sensor's active area, or else some of your load will bypass the active area, leading to systemic errors. And you'd better radius the edges, lest you create a stress concentration right there.